You can use the Kubernetes Operator to deploy MongoDB Search and Vector Search across multiple Kubernetes clusters for an external MongoDB deployment. In this case the Kubernetes Operator just manages mongot, not the external MongoDB deployment. Your MongoDB deployment is external to the Kubernetes Operator. It can run:
Outside Kubernetes, such as on virtual machines
In Kubernetes clusters, but not managed by the Kubernetes Operator
The Kubernetes Operator runs in a single, central cluster and deploys a mongot group into each Kubernetes cluster that you list in the spec.clusters field of the MongoDBSearch custom resource. Each mongot group syncs only from the MongoDB members that you map to it. You control this mapping with member tags, which you set later in this procedure.
Select your configuration to filter the installation procedure for your deployment.
Note
Multi-cluster deployments require MongoDB Enterprise Edition and an external MongoDB deployment. Multi-cluster deployments in which the Kubernetes Operator manages the MongoDB deployment aren't currently supported as source to the multi-cluster Search deployment. See the Supported Topologies.
Next Steps
After mongot pods are running across all member clusters, you can:
Connect to MongoDB Search and Vector Search and run queries. See Use MongoDB Search and Vector Search.
Review
MongoDBSearchconfiguration options. See MongoDB Search and Vector Search Settings.